Unsteady Characteristics of Forward Multi-Wing Centrifugal Fan at Low Flow Rate

Round 1

Reviewer 1 Report

Stead and unsteady CFD computations using LES and structured grid technology were performed for the centrifugal fan at a small flow rate. The flow separation and pressure fluctuation phenomena were predicted and analyzed based on the computations. Q criteria were used for flow field evaluations. Spectrum analyses were employed for the pressure distribution and kinetic energy of internal flow. Authors observed the separation on the suction surface and trailing edge of the blade.

LES and hybrid method are now very popular way to study the complex flow phenomena. It is very meaningful to apply the advanced methods to the really applications like centrifugal fans and explore details of physics so as to solve issues. In this manuscript, authors are trying to identify reasons for poor performance of centrifugal fan with low flow rate.

The paper is well organized. The analyses are consistent with results and assessment appears reasonable.

Some comments the reviewer wants to share with authors for the improvement of the paper:

For the fan’s flow condition, please explicitly state the tip Mach number of the blade. SIMPLE algorithm is pressure based and good for the incompressible flow. That is another reason that the tip Mach number is a key factor to show the flow characteristics and suitability of the algorithm utilized. LES is quite strict for the grid quality and size; otherwise the advantage of the LES cannot be achieved. Based on the grid sized used in your work, I am not sure the quality of the grid. Please add one RANs solution and compare with your LES solution claimed in this work. Pressure prediction probably is most easy one to match the experiment. The pressure agreement cannot guarantee other flow properties’ prediction. Therefore, please add the Cp distributions along various radial locations. On the other hand, Cf distributions can tell something. Please add the Cf distribution on the both of the pressure side and suction side with the comparison to the experiment. For the pressure pulsation analyses, listing all fluctuations in Figure 14 at various locations doesn’t seem to be useful to help people understand the flow property. Please provide the conclusive analyses. For all the plots, the labels and legend’s scales are way to0 small. Please adjust them accordingly.

Author Response

Reviewer1:

Responses to Comments and Suggestions for Authors

1.1 For the fan’s flow condition, please explicitly state the tip Mach number of the blade. SIMPLE algorithm is pressure based and good for the incompressible flow. That is another reason that the tip Mach number is a key factor to show the flow characteristics and suitability of the algorithm utilized.

Response: Thank you for your affirmation and valuable comments. Corrections have been made.

The tip Mach number of the blade is given in our paper., Mach number equals 0.03. When the Mach number is less than 0.3, the fluid state is a subsonic incompressible gas.

 1.2 LES is quite strict for the grid quality and size; otherwise the advantage of the LES cannot be achieved. Based on the grid sized used in your work, I am not sure the quality of the grid.

Response: Thank you for your prompt. The quality of the grid used in the calculation has been given.

The LES has a high demand for grid. Therefore, the computational domain of centrifugal fan is divided into structural grids. The mesh quality in the left entrance domain is at a range of 0.735 to 0.996. The mesh quality in the right entrance domain is at a range of 0.655 to 0.997. The mesh quality in the impeller domain is between 0.875 and 0.997. The mesh quality in the volute domain is between 0.612 and 0.999. The mesh quality in the left entrance domain is at a range of 0.735 to 0.996. The grid quality of the exit domain is at a range of 0.95 to 1.

 1.3 Please add one RANS solution and compare with your LES solution claimed in this work

Response: Thank you for your valuable suggestions. The paper is further modified and the difference of LES and RANS models is explained. The unsteady calculations of centrifugal fan are also carried out using LES、RNG k-ε、realizable k-ε and standard k-ε models in view of same mesh system at flow rate of Q/Qn=0.031. The same time step is all adopted for unsteady simulation and the results at the same time are analyzed. The difference of Q value distribution is represented at the Z=0.015m plane in Figure 1. The LES model clearly shows more details of the internal vortex structure. On the contrary, the results calculated by RANS models have limitations on small-scale vortices. LES model is very useful for most studies of unsteady flow phenomena. when the small vortices are easily dissipated in the calculation process, the accuracy of simulation can be improved by using the model instead of the small vortices. Therefore, the following explanations of unsteady flow are all based on the numerical results using LES model .

Figure 1 Q value distribution between LES and RANS at the same computing time t=0.0125s

1.4 Pressure prediction probably is most easy one to match the experiment. The pressure agreement cannot guarantee other flow properties’ prediction. Therefore, please add the Cp distributions along various radial locations. On the other hand, Cf distributions can tell something. Please add the Cf distribution on the both of the pressure side and suction side with the comparison to the experiment.

Response: Thank you for your valuable advice. The static pressure coefficients of different radial position monitoring points have been given in Figure 2.

(a)

(b)

(c)

(d)

(e)

(f)

Figure 2 Static pressure coefficient of monitoring points at different radial positions

1.5 For the pressure pulsation analyses, listing all fluctuations in Figure 14 at various locations doesn’t seem to be useful to help people understand the flow property. Please provide the conclusive analyses. For all the plots, the labels and legend’s scales are way too small. Please adjust them accordingly. 

Response: Thank you for your affirmation and valuable comments. Analysis of static pressure pulsation at different monitoring points is re-described and all drawing are update.

Author Response File: Author Response.docx

Reviewer 2 Report

The topic of choice and work that you did has a sound structure to it. This means that your progress and obtained results look promising and valid. That being said, both the post-processing of the results and presentation of the paper must go under significant revisions and improvements.

In my opinion the post-processing and validation of steady state results requires improvement and don't seem to be promising as they are now. Hence, I did not review the unsteady results at all.

See the provided comments in the attached PDF. 

Comments for author File: Comments.pdf

Author Response

1 Reviewer2

Responses to Comments and Suggestions for Authors

1.1 Be more specific about the application of these type of machines. Why they are important to go under study? What are their application? Why you need to address their limitation?

Any turbomachine operating at significantly lower or higher volumetric flow rate compared to its BEP would face operation issues. You should be specific with the term "small flow rate"! Compare the normal operating conditions of these type of machines and possibly express it in terms of non-dimensional numbers.

Response: Thank you for your valuable suggestions. Follow your advice and give a further description of the introduction.

Forward multi-wing centrifugal fan is widely used in air conditioning industry because of a great deal of salient virtues, for instance high pressure coefficient, small size coefficient, and low noise. The performance and noise of centrifugal fans have attracted more and more attention. Nowadays, the rising attentions are paid to the performance and noise of centrifugal fan. Therefore, it is very important for the optimization of centrifugal fan to master the internal flow law of centrifugal fan. This will not only improve the competitiveness of centrifugal fans, but also benefit the development of the whole fan industry.

1.2 As the literature review you are just mentioning multiple works using different numerical model approach（from RANS to LES）.You need to cite related work to your approach and conclude what they conclude and what they did not.

In the end you need to clearly point out the message and goal of the current work and more detailed description of the paper structure.

Response: Thank you for your valuable suggestions. Corrections have been made

The steady internal flow in centrifugal fan is studied by implementing Reynolds-averaged Navier-Stokes Equations method (RANS). As a general rule, RANS model solves the average information of the flow field. The unsteady turbulence problem is transformed into a steady one, which is sufficient to study the aerodynamic performance of centrifugal fans.

The RNG k-ε mainly used for the steady calculation in this paper. The e equation in RNG turbulence model improves the ability to simulate high strain flow. The RNG turbulence model takes turbulent vortices into account and improves the accuracy in this respect and provides an analytical formula considering the viscous flow of Landmine Nobel Number. This makes RNG turbulence model more accurate and reliable so that swirling flow with moderate intensity and low Reynolds number flow will be well predicted.

 1.3 This subsection is redundant! It is basically copy/paste of the help manual of every commercial software.

It is highly suggested to avoid using equations and basic definitions.

If you have used any special models or modified the currently available models for any reason, then describe it here. How and why you modified or use the model.

As an example why the Q Criterion was used? How it would hurt the CFD result if not used?

As you are citing there are 5+ and even more work on this topic and criterion.

This is how your work would add a value to community, not re-writing the RANS or LES governing equations.

Response: Thank you for your valuable comments. Equations and basic definitions have been deleted. The reasons for using RANS, LES and Q criterion are also explained in detail in this paper.

The steady internal flow in centrifugal fan is studied by implementing Reynolds-averaged Navier-Stokes Equations method (RANS).As a general rule, RANS model solves the average information of the flow field. The unsteady turbulence problem is transformed into a steady one, which is sufficient to study the aerodynamic performance of centrifugal fans.

The RNG k-ε mainly used for the steady calculation in this paper. The e equation in RNG turbulence model improves the ability to simulate high strain flow. The RNG turbulence model takes turbulent vortices into account and improves the accuracy in this respect and provides an analytical formula considering the viscous flow of Landmine Nobel Number. This makes RNG turbulence model more accurate and reliable so that swirling flow with moderate intensity and low Reynolds number flow will be well predicted.

The RANS model is usually used to calculate the time averaged flow field. The RANS model essentially changes the actual problem, abandons the simulation of unsteady turbulence information, and only seeks the flow results in the average time. Hence, it is impossible to simulate unsteady turbulent flow accurately. Nevertheless, the large eddy simulation is a turbulent model to simulate unsteady flow. It solves the instantaneous information of flow field.

Unsteady calculations of centrifugal fan are also carried out using LES、RNG k-ε、realizable k-ε and standard k-ε models in view of same mesh system at flow rate of Q/Qn=0.031.The same time step is all adopted for unsteady simulation and the results at the same time are analyzed. The difference of Q value distribution is revealed at the Z=0.015m plane in Figure 1. The LES model clearly shows more details of the internal vortex structure. On the contrary, the results calculated by RANS models have limitations on small-scale vortices. LES model is very useful for most studies of unsteady flow phenomena. Because the small vortices are easily dissipated in the calculation process, the accuracy of simulation can be improved by using the model instead of the small vortices. Therefore, the following explanations of unsteady flow are all based on the results of LES model calculation.

Figure 1 Q value distribution between LES and RANS at the same computing time t=0.0125s

 The criteria that is used to judge vorticity mainly include the Q criterion and vorticity criterion. The application of vorticity criterion requires a given threshold, but different flow thresholds may be different, and different thresholds will result in different results. The Q criterion reflects a balance between rotation and deformation of a fluid cluster in the flow field. Q > 0 reflects the dominance of rotation in flow. At the same time, the Q criterion has no threshold selection, which is more objective than the vorticity criterion.

1.4All of the figures in this paper are not clear!

Elaborate and have chart of the boundary conditions and places of them.

Elaborate more on used mesher and mesh setting.

Round the total mesh element number (i.e. 4.3[M]instead of 4,328,863).

Response: Thank you for your valuable suggestions. All of the figures in this paper are redrawn.

The computational domain grids are reorganized in Figure 2. And writing method of the total mesh element number has been changed.

 (a)

(b)

(c)

(d)                                (e)

Figure 2 Mesh of compute domain:

(a)the overall structure mesh of centrifugal fan; (b)the mesh of volute;(c) the mesh of impeller;(d) the boundary layer of volute;(e) detail of the mesh near blade

1.5 No explanation on how you are modeling the rotor and stator sections and numerical interpolation between them (i.e. RRF, Sliding mesh, Mixing Plane, Frozen Rotor)?

Response: Thank you for your valuable suggestions.

It is necessary to set frozen rotor in the steady simulation, and sliding mesh is applied for unsteady simulation between dynamic and static domains.

1.6 Quantify the error values between CFD and experiment.

I am not seeing any citation or explanation on experiment!

Why CFD and experiment are closely matching at higher vol. flow rates but not at lower values (maybe you should tell reader that you observed this Intro).

Response: Thank you for your valuable suggestions.

This performance experiment of centrifugal fan was tested in Zhejiang Yilida aerodynamics and acoustic laboratory. Figure 3 shows the exterior and interior of the semi-anechoic room in the laboratory. The performance and noise data of centrifugal fan in semi-anechoic room can be tested at the same time. Installation method of centrifugal fan and instrument are also shown in Figure 3.

(a) Laboratory appearance and internal Scene

(b) Installation Method of Centrifugal Fan

(c) Speed Sensor and Instrument Cabinet

(d) Two-dimensional map of semi- anechoic room

Figure 3 Laboratory test system

1.7 What is the volumetric flow rate you ran CFD for? Where are you located on the curve?

Here is the suggestions:

Run the CFD at higher vol. flow rate where CFD and experiment result are matching. Go through the result and show the flow field under optimum conditions. Run CFD on lower vol. flow rates. Present your result clearly and normalize you contours with the baseline(i.e. optimum conditions.)

Response: Thank you for your valuable suggestions.

In this paper, the main data of analysis is the result of flow rate Q/Qn=0.031.

The whole performance curve is simulated and compared by CFD numerical simulation according to the working conditions in the experiment. And steady internal contours with optimum condition and high flow rate are added in this paper.

Other details have also been changed in this paper.

Author Response File: Author Response.docx

Reviewer 3 Report

I recommend that you provide the software name or program creators and the language of the software. It is advisable to specify the operating system in which computer simulations were performed.

Author Response

Review 3:

Responses to Comments and Suggestions for Authors

I recommend that you provide the software name or program creators and the language of the software. It is advisable to specify the operating system in which computer simulations were performed.

Response: Thank you for your valuable comments. The software name and the operating system are simply described.

A numerical simulations based on the finite-volume numerical method using FLUENT 16.0 software were carried out to analyse internal flow of centrifugal fan under steady and unsteady computations, respectively. All numerical calculations were carried out using a DELL computer of including two CPU(2.30GHz) under the Windows operating system

 Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comparing with the previous version, authors did a lot of grammars and reorganizing work. However, authors didn’t address most of my questions in my previous comments.

Again, since this is mainly a CFD work trying to explore the physics of the flow field for the small centrifugal fan, the experiment data is used for cross-reference the accuracy of the CFD simulation. However, the pressure distribution is the only experimental data available and it is not enough to demonstrate the capability of the simulation especially for the LES simulation. Also, for the simulation results, authors didn’t conclude the new in-depth observations for the simulation. In my opinion, this work needs to be further investigated.

Author Response

Please see the Attached file.

Author Response File: Author Response.docx